Researchers propose a framework called Traceable Trust to ensure the reliability of artificial intelligence in bioscience. They argue that AI outputs should be subject to a review process before guiding laboratory actions. The framework assesses evidence supporting AI predictions, capability claims, delegated agency, and decision thresholds. Three case studies demonstrate how trust can be documented when AI shapes scientific work. --- Why it matters: This matters because AI is increasingly used in bioscience research, but its outputs often lack transparency and accountability. Traceable Trust provides a framework for evaluating the reliability of AI predictions and ensuring that they are actionable. Source: https://arxiv.org/abs/2608.17997